Clarksville.

library(aqp)
library(soilDB)

# get lab / morphologic data
x <- fetchKSSL(series = 'clarksville', returnMorphologicData = TRUE, simplifyColors = TRUE)

# extract SoilProfileCollection
s <- x$SPC

# remove horizons that are missing moist colors
s <- subsetHz(s, !is.na(m_hue) & !is.na(m_value) & !is.na(m_chroma))

# remove profiles with missing horizons due to above steps
s <- HzDepthLogicSubset(s)

# keep only profiles with > 2 horizons
idx <- which(profileApply(s, nrow) > 2)
s <- s[idx, ]

# re-assemble Munsell color notation for moist color
s$color <- sprintf("%s %s/%s", s$m_hue, s$m_value, s$m_chroma)

One approach: extract the first n profiles having > 5 horizons, and first 6 horizons.

# filter to just those profiles with > 5 horizons
idx <- which(profileApply(s, nrow) > 5)

# take the first n profiles, all horizons
n <- 6
w <- s[idx[1:n], 1:6]

# check
par(mar=c(0,0,0,0))
plotSPC(w, color = 'moist_soil_color', cex.names = 0.75, width = 0.35, name.style = 'center-center', plot.depth.axis = FALSE, hz.depths = TRUE, fixLabelCollisions = TRUE, hz.depths.offset = 0.06)

Less restrictive: first n profiles, all horizons.

n <- 6
w <- s[1:n, ]

# check
par(mar=c(0,0,0,0))
plotSPC(w, color = 'moist_soil_color', cex.names = 0.75, width = 0.35, name.style = 'center-center', plot.depth.axis = FALSE, hz.depths = TRUE, hz.depths.offset = 0.06)

Regularize horizons / colors to k rows.

regularizeColors <- function(i, k = 6) {
  .col <- colorRampPalette(na.omit(i$moist_soil_color), space = 'Lab', interpolate = 'spline')(k)
  .munsell <- col2Munsell(t(col2rgb(.col)))
  .munsell <- sprintf("%s %s/%s", .munsell$hue, .munsell$value, .munsell$chroma)
  
  return(list(col = .col, munsell = .munsell))
}

mm <- profileApply(w, simplify = FALSE, FUN = regularizeColors, k = 6)

Make a k row by n column grid of soil colors from above.

m <- lapply(mm, '[[', 'col')
m <- do.call('cbind', m)

# m <- matrix(m$moist_soil_color, nrow = 6, ncol = 5, byrow = FALSE)

par(mar=c(0,0,0,0))
plot(1, 1, type='n', axes=FALSE, xlab='', ylab='', ylim=c(6.5, 0.5), xlim=c(0.5, n+0.5), asp = 1)

# vectorized functions are the best
rect(xleft = col(m) - 0.4, ybottom = row(m) + 0.4, xright = col(m) + 0.4, ytop = row(m) - 0.4, col = m, border = NA)

Annotate with Munsell colors.

.munsell <- lapply(mm, '[[', 'munsell')
.munsell <- do.call('c', .munsell)

par(mar=c(0,0,0,0))
plot(1, 1, type='n', axes=FALSE, xlab='', ylab='', ylim=c(6.5, 0.5), xlim=c(0.5, n+0.5), asp = 1)

# vectorized functions are the best
rect(xleft = col(m) - 0.4, ybottom = row(m) + 0.4, xright = col(m) + 0.4, ytop = row(m) - 0.4, col = m, border = NA)

text(x = col(m), y = row(m), labels = .munsell, col = invertLabelColor(m), cex = 0.66, font = 2)

This time with rounded corners

library(berryFunctions)

par(mar=c(0,0,0,0))
plot(1, 1, type='n', axes=FALSE, xlab='', ylab='', ylim=c(6.5, 0.5), xlim=c(0.5, n+0.5), asp = 1)

# ugh
col.m <- col(m)
row.m <- row(m)

# not vectorized
for(i in seq_along(m)) {
  roundedRect(xleft = col.m[i] - 0.4, ybottom = row.m[i] + 0.4, xright = col.m[i] + 0.4, ytop = row.m[i] - 0.4, col = m[[i]], border = NA, rounding = 0.25)
}

Arbitrary collection of profiles

library(sharpshootR)
data("OSDexamples")

s <- OSDexamples$SPC
s$color <- sprintf("%s %s/%s", s$hue, s$value, s$chroma)

s$moist_soil_color <- s$soil_color

mm <- profileApply(s, simplify = FALSE, FUN = regularizeColors, k = 8)

m <- lapply(mm, '[[', 'col')
m <- do.call('cbind', m)

n <- length(s)

# init plot region
par(mar=c(0,0,0,0))
plot(1, 1, type='n', axes=FALSE, xlab='', ylab='', ylim=c(6.5, 0.5), xlim=c(0.5, n+0.5), asp = 1)

# keep track of column and row indices for simpler plotting
col.m <- col(m)
row.m <- row(m)

# roundedRect() not vectorized
for(i in seq_along(m)) {
  roundedRect(
    xleft = col.m[i] - 0.4, 
    ybottom = row.m[i] + 0.4, 
    xright = col.m[i] + 0.4, 
    ytop = row.m[i] - 0.4, 
    col = m[[i]], 
    border = NA, 
    rounding = 0.25
  )
}